N. J. Pramod Dhinakar 1, G. Kishore Kumar 2, M. Raghavendra 3

A Review on Mobile Cloud Computing N. J. Pramod Dhinakar 1, G. Kishore Kumar 2, M. Raghavendra 3 1 Asst. Professor, Dept. of Information Technology, RGM College of Engg. & Tech., Nandyal, Kurnool Dt. 2 Asst. Professor, Dept. of Information Technology, RGM College of Engg. & Tech., Nandyal, Kurnool Dt 3 Asst. Professor, Dept. of Information Technology, RGM College of Engg. &Tech., Nandyal, Kurnool Dt 1 it.pramod1211@gmail.com 2 kishorgulla@yahoo.co.in 3 raghusavoli@gmail.com Abstract Mobile Cloud Computing (MCC) is an infrastructure of data storage and data processing which happens outside of the mobile devices. Mobile Cloud Computing is the advanced distributed mobile computing technology that is recently gaining ground to augment computational capabilities of resource constraint mobile devices using cloud-based resources. MCC provides unrestricted functionalities, rich mobile computational resources, computational technology, platforms, heterogeneous environment, and business opportunities for cloud computing providers, mobile users and network operators. In this paper, we present can overview of MCC, its definition, motivation, and organization of MCC building blocks. Keywords Mobile Cloud Computing, Mobile Cloud Computing Architecture, Cloud Computing. I. INTRODUCTION Cloud Computing is the metaphor for the internet. Cloud Computing provides virtually abundant amount of dynamic resources, on demand computing capacity, large amount of storage, services to individuals and businesses in the form of heterogeneous environment and autonomous services. Therefore cloud computing extends to mobile devices. The popularity and availability of mobile devices, especially smartphones are creating tense dependency that people do not leave home without them. However, smartphones miniature nature, lightness, and mobility traits impose severe limitations on their processing capabilities, battery lifetime, storage capacity, and visualization power (e.g., screen size and rendering capability) impeding execution of resourceintensive computations and bulky data storage on smartphones. Cloud Computing has been widely recognized as the next generation s computing infrastructure. Cloud Computing offers some advantages by allowing users to use infrastructure (e.g., servers, networks, and storages), platforms (e.g., middleware services and operating systems), and software (e.g., application programs) provided by cloud providers (e.g., Google, Amazon, and Salesforce) at low cost. In addition, Cloud Computing enables users to elastically utilize resources in an on-demand fashion. As a result, mobile applications can be rapidly provisioned and released with the minimal management efforts or service provider s interactions. Mobile Cloud Computing (MCC) is introduced as an integration of Cloud Computing into the mobile environment. Mobile Cloud Computing brings new types of services and facilities for mobile users to take full advantages of cloud computing. II. MOBILE CLOUD COMPUTING In this section, we present motivation to extend mobile devices and present MCC definition. A. Motivation The MCC motivation lies in intrinsic deficiencies of mobile devices and realization of ever increasing computing requirements of mobile users. The miniature nature and mobility requirement of mobile devices enforce significant constraints on their CPU, RAM, storage and battery. Mobile device manufacturers are endeavouring to enhance computing capabilities of mobile devices by employing energy efficient multi-core processors, large fast RAM, massive low-overhead storage and high charge density (long life) battery. Alternatively, researchers use the concept of on-demand, rich computing resources of cloud computing to improve mobile devices limitations and fulfil limitless users demand that breeds the state-of-the-art Mobile Cloud Computing (MCC). B. Definition MCC is a rich mobile computing technology that leverages unified elastic resources of varied clouds and network technologies toward unrestricted functionality, storage, and mobility to serve a multitude of mobile devices anywhere, anytime through the channel of Ethernet or Internet regardless of heterogeneous environments and platforms based on the pay-as-you-use principle We can also define MCC as Mobile Cloud Computing at its simplest, refers to an infrastructure where both the data storage and the data processing happen outside of the mobile device. Mobile cloud applications move the computing power 352

and data storage away from mobile phones and into the cloud, B. Software bringing applications and mobile computing to not just The software building block of MCC comprises smartphone users but a much broader range of mobile augmentation protocols and solutions to efficiently leverage subscribers. cloud-based resources to mitigate shortcomings of mobile devices. III. BUILDING BLOCKS OF MOBILE CLOUD COMPUTING In this section, we devise classification of MCC main building blocks and explain them accordingly. Mobile Cloud Computing is studied from two aspects hardware and software. In every MCC system, hardware building blocks provide a rich mobile computation platform that can be employed by varied software programs. Our devised classification is illustrated in Fig. 1 and explained as follows. A. Hardware Hardware infrastructures, including heterogeneous resource-constraint mobile devices, cloud based resources, and networking infrastructures are physical building blocks of MCC. Heterogeneity in MCC is inherited from mobile and cloud computing technologies and is intensified in the existence of a multitude of dissimilar devices, infrastructures, technologies, and features. 1) Mobile Devices: MCC is fraught by multitude of heterogeneous battery-operating wirelessly-connected mobile devices (e.g., Smartphones, tablets, and wearable computers) that feature varied limited computing capabilities. 2) Cloud-based Resources: Computing entities that are built based on cloud computing technologies and principles (e.g., elasticity and pay-as-you-use) are called cloud-based resources. Four types of cloud-based resources are identified in as Distant Immobile Clouds (DIC), Proximate Immobile Clouds (PIC), Proximate Mobile Clouds (PMC), and Hybrid (H). 3) Networking Infrastructures: Efficient, reliable, and high performance networking in MCC necessitates deployment of both wired and wireless networking technologies and infrastructures. Although mobile devices perform only wireless communications, immobile cloud-based resources require wired communication to transmit digital contents to different computing devices in a reliable and high speed medium. Cloud-based Mobile Augmentation (CMA): CMA is the advanced mobile augmentation model that leverages cloud computing technologies and principles to increase, enhance, and optimize computing capabilities of mobile devices by executing resource-intensive mobile application components in the resource-rich cloud-based resources. 1) Computation Offloading: It is the process of identifying, partitioning, and migrating resource-intensive components of mobile applications to cloud-based resources. Identifying intensive components and partitioning can take place in three different approaches of static, dynamic, and hybrid. Static partitioning is a one-time process of identifying and partitioning the intensive components of mobile application at design and development time. The benefit of static partitioning is that it does not impose runtime overhead on a mobile device and once the application is partitioned, the same partitions can be used for infinite executions. However, static partitioning is not adaptable to the environment changes and its dynamism. On the contrary, in dynamic partitioning the identification and partitioning take place at runtime to better meet dynamism of the MCC environment. The challenge in dynamic partitioning is the excessive overhead of identifying intensive tasks, monitoring the environment, partitioning the application, and offloading the components. The third approach is to use a hybrid model where some part of the application is partitioned at design time and some at runtime to mitigate the partitioning overhead and adapt to the environmental changes. 2) Multi-tier Programming: It is proposed and employed to alleviate the overhead of code partitioning and offloading by building loosely coupled applications that 353

Fig 1. Classification of Mobile Cloud Computing Building Blocks perform resource-intensive computations (often web services) in the remote resources and minimize mobile-side computations. Resource-intensive computations are always available in the remote servers to be called for execution. Thus, the overhead of identifying, partitioning, and migrating tasks from mobile device to remote resources is omitted. Upon successful execution of the intensive tasks, the results are synchronized with the native components in the mobile device. In this model, only data is transmitted to the remote resources and codes are not migrated from the mobile device. Thus, the transmission overhead is significantly reduced. 3) Live Cloud Streaming: It is another approach that aims to augment mobile devices by performing the entire computations outside the mobile device. Results are delivered to users as pre-compiled screenshots streaming to the mobile device lively. This approach requires low latency, high throughput, reliable wireless network that is challenging to establish using current technologies. For instance, www.onlive.com delivers fascinating games to the mobile users via live cloud streaming. 4) Remote Data Managing Solutions: The Dropbox virtually expand mobile storage by storing users digital contents in the cloud-based resources. Parallel to growth in computing, digital data are sharply increasing that demand huge space on mobile devices which further encumbers mobile device adoption and usability. Although cloud storages enhance storage deficiency of mobile devices and improve data safety, trust and data security and privacy prevent users to intuitively leverage remote storages. IV. MOBILE CLOUD COMPUTING ARCHITECTURE There are various MCC proposals have been investigated over the last few years developing four major complementing architectures for MCC, which are briefly discussed in this section. Table 1 presents major characteristics of four MCC architectures. It is noteworthy that these architectures are applicable to all cloud-based augmentation models described in Fig. 1. A. Mobile Distant-Immobile-Cloud Computing (MDICC) A general abstract architecture for a typical MCC system comprises of a mobile device that consumes computing resources of a computing entity using a typical offloading framework like MAUI via a bidirectional wireless link. The first proposed architecture for MCC is depicted in Fig. 2 (a) where the mobile user is consuming computational resources of public clouds using the channel of Internet. 354

TABLE I MAJOR CHARACTARISTICS OF VARIED MOBILE CLOUD COMPUTING ARCHITECTURES Characteristics MDICC MPICC MPMCC HMCC Architecture Client-Server Client-Server/Peer-to-Peer Heterogeneity High High Low Medium WAN Latency High Medium Low Medium Resource Elasticity High Medium Low High Resource Multiplicity Low Medium High High Resource Availability High Low Medium High Mobility Implication Medium (client-side mobility) High since both can move Utilization Cost Low Medium High Medium Security & Privacy High Medium Low High Trust High Medium Low High Computational tasks in this model are executed inside the DIC resources (i.e., public cloud service providers such as Amazon EC2) and the results are sent back to the mobile client. The main advantages of DIC are high computational capabilities, resource elasticity, and relatively high security. High computing capabilities and elasticity of DIC minimize remote computational time and conserve mobile battery. The utilization cost of DIC data centres is the least possible cost considering the ultimate goal of cloud computing to reduce the computing costs. Heterogeneity also imposes excess overhead by employing handling techniques such as virtualization, semantic technology, and middleware systems. Moreover, DIC are coarse location granular resources (meaning they are few in numbers located far away from majority of mobile service consumers) and are intrinsically immobile computing resources. Thus, exploiting DIC for augmenting mobile devices originates long WAN latency due to manifold intermediate hops and high data communication overhead in the intermittent wireless networks. Long WAN latency degrades the application execution performance and wastes limited mobile battery. Moreover, service consumers mobility on one hand and lack of clouds mobility on the other hand intensify WAN latency, and degrade effectiveness and efficiency of MCC solutions. In fact WAN latency will likely remain in wireless communications for a long time despite of significant improvements in data transmission speed and network bandwidth. The main delaying factor in WAN latency is the processing delay at each intermediate hop to perform tasks such as decompression, decryption, security checking, virus scanning, and routing for each packet. Thus, the larger is the number of hops, the longer is the WAN latency. B. Mobile Proximate-Immobile-Cloud Computing (MPICC) To mitigate the impacts of long WAN latency, researchers endeavour to access computing resources with least number of intermediate hops and propose alternative architecture for MCC depicted in Fig. 2 (b). Hence, mobile devices utilize computing resources of desktop computers in nearby public places such as coffee shops and shopping malls. Instead of travelling through numerous hops to performing intensive computations in DIC, tasks are executed inside the one-hop distance public computers in vicinity (called PIC). PICs are medium location-granular (compared to the coarse grain resources, PICs are more in number and are located nearer to mobile service consumers) and feature moderate computational power that provide less scalability and elasticity. Moreover, employing computing resources of PICs holds several implications, particularly service provider and consumers security and privacy, isolating a computer s host OS from a guest mobile OS, incentivizing computer owners and encourage them to share resources with nearby mobile devices, lack of on-demand availability (shops are open in certain hours and days), and lack of PIC mobility that demand future research. C. Mobile Proximate-Mobile-Cloud Computing (MPMCC) The third MCC architecture depicted in Fig. 2 (c) is recently proposed to employ a cloud of nearby resourcerich mobile devices (i.e., PMCs) that are willing to share resources with proximate resource-constraint mobile devices. Rapidly emerging popularity and ever-increasing multiplicity of contemporary mobile devices, especially 355

Fig 2. Illustrative View of Varied Mobile Cloud Computing Architectures smartphones and tablets realize the vision of building PMC. Two different computing models are feasible which are peer-to-peer and client-server. In peer-to-peer, service consumers and providers can directly communicate with each other to negotiate and initiate the augmentation. However, service consumer needs to perform energyconsuming node discovery task to find appropriate mobile service provider in vicinity. Moreover, peer-to-peer systems are likely vulnerable to fraudulent service providers that can attack the service consumer device and violate its privacy. The alternative MPMCC communication model is arbitrated client-server in which mobile client communicates with a trusted arbitrator and request for the most reliable and appropriate proximate node. The arbitrator can keep track of different service providers and perform security monitoring on mutual communications. The crucial advantages of exploiting such resources are negligible service provider and consumers heterogeneity, resource pervasiveness, and short WAN latency. In either client-server or peer-to-peer models, the number of intermediate hops are small due to service provider and consumer s vicinity. The resource pervasiveness of PMCs enables execution of resourcehungry tasks anytime anywhere (either in an ad-hoc ecosystem or infrastructure environment where Mobile Network Operators (MNO) like Verizon can manage the process). Therefore, diminutive WAN latency, and high multiplicity and ubiquity of mobile service providers establish a solid ground to leverage such ubiquitous resources in high latency-sensitive, low security-sensitive mobile computational tasks. Although resource multiplicity in this architecture is high, scalability and elasticity is limited due to constraint computing power of individual mobile devices. Mobility management is another challenging feature of this architecture. Unrestricted mobility of mobile service providers and mobile service consumers in this architecture significantly complicates seamless connectivity and mobility, and noticeably degrade efficiency of augmentation solutions. When a user (either service consumer or provider) starts moving across heterogeneous wireless networks with dissimilar bandwidths, jitters, and latencies, the varying network throughputs, increasing mobile-cloud distance, and frequent network disconnections increase WAN latency and directly impact on application response time and energy efficiency. However, the utilization cost of PMCs is likely higher than 356

other immobile resources due to ubiquity and negligible Computing, IEEE Transactions on Consumer Electronics, Feb 2014, pp. 1-9 latency, though they feature finite computing resources. Other shortcomings of this model are security, privacy, and [2] Abolfazli, S., Sanaei, Z., Ahmed, E., et al., 2014. Cloud-based data safety in mobile devices. Mobile devices are not safe Augmentation for Mobile devices: Motivation, Taxonomies, and Open Challenges. IEEE Communications Surveys & Tutorials, to store user data because they are susceptible to physical 16(1),pp.337 368 damage, robbery, hardware malfunction, and risk of loss. D. Hybrid Mobile Cloud Computing (HMCC) Each of the mentioned three architectures features cons and pros that encumber optimal exploitation for efficient mobile computation augmentation. Researchers in demonstrate feasibility of consolidate various resources to build a HMCC model (see Fig. 2 (d)) that mitigates deficiencies of pervasive systems and promotes strength towards optimal CMA. SAMI is a multi-tier infrastructure that convergences public clouds, MNOs, and MNOs trusted dealers to optimize the augmentation of resourceconstraint mobile devices. In the core of the multi-tier infrastructure, a resource scheduler program needs to evaluate each computational task and allocate appropriate resource(s) from the silo of heterogeneous cloud-based resources that optimally meet computational needs and fulfil user s QoS requirements (like cost, security, and latency). However, considering challenges, especially seamless mobility, significant MCC heterogeneity, wireless network intermittency, and current wireless networking, developing an optimal generic scheduler is a non-trivial task. Moreover, increasing number of mobile service consumers and hybrid cloud-based resources increase system complexity and complicate management and maintenance. Lightweight resource discovery and scheduling algorithms are essential for this MCC architecture. V. CONCLUSION Mobile Cloud Computing (MCC) is the advanced distributed mobile computing technology aiming to alleviate resource deficiency of multitude of heterogeneous resource constraint mobile devices by performing resourceintensive computations inside cloud-based resources. Mobile augmentation solutions are highly influenced by the employed cloud-based resources. Granularity (resource multiplicity and proximity), computational capability, mobility, and heterogeneity between mobile device and resources are major influential resource properties that impact on augmentation performance and needs consideration in design and development of imminent augmentation solutions. [3] Abolfazli, S., Sanaei, Z., Shiraz, M., et al., 2012. MOMCC: Marketoriented architecture for Mobile Cloud Computing based on Service Oriented Architecture. Proceedings IEEE Mobile Cloud Computing Workshop, Beijing,pp. 8 13 [4] Abolfazli, S., Sanaei, Z., Gani Abdullah, et al., 2013. Rich Mobile Application: Genesis, Taxonomy, and Open Issues. Journal of Network and Computer Applications, Vol 40, pp. 345-362 [5] Clark, C. et al., 2005. Live Migration of Virtual Machines. Proceedeings Symposium Networked Systems Design and Implementation,Boston,pp. 273 286 [6] Cuervo, E. et al., 2010. MAUI: Making Smartphones Last Longer With Code Offload. Proceedings ACM Conference Mobile Systems, Application, Services, San Francisco,pp. 49 62 [7] Sanaei, Z., Abolfazli, S., Khodadadi, T., et al., 2013. Hybrid Pervasive Mobile Cloud Computing: Toward Enhancing Invisibility. INFORMATION-An international interdisciplinary journal, 16(11),pp.1 12 [8] Satyanarayanan, M. et al., 2009. The Case for VM-Based Cloudlets in Mobile Computing. IEEE Pervasive Computing,8(4),pp.14 23 [9] http://www.mobilecloudcomputingforum.com/ [10] M. Satyanarayanan, Mobile computing: the next decade, in Proceedings of the 1st ACM Workshop on Mobile Cloud Computing & Services: Social Networks and Beyond (MCS), June 2010. [11] J. Flinn, S. Park, M. Satyanarayanan, Balancing performance, energy, and quality in pervasive computing, in: Proceedings of the 22nd International Conference on Distributed Computing Systems, 2002, IEEE, 2002, pp. 217 226. [12] Saeid abolfazli* a, zohreh sanaeia, mohammad hadi sanaeia, mohammad shojafarb, abdullah gania, Mobile cloud computing: the state-of-the-art, challenges, and future research. References [1] Abolfazli, S., Sanaei, Z., Alizadeh, M., et al., 2014. An Experimental Analysis on Cloud-based Mobile Augmentation in Mobile Cloud 357